Effect of ankle angles on the soleus H-reflex excitability during standing

Abstract

This study investigated effects of ankle joint angle on the H-reflex excitability during loaded (weight borne with both legs) and unloaded (full body weight borne with the contralateral leg) standing in people without neurological injuries. Soleus H-reflex/M-wave recruitment curves were examined during upright standing on 3 different slopes that imposed plantarflexion (−15°), dorsiflexion (+15°), and neutral (0°) angles at the ankle, with the test leg loaded and unloaded. With the leg loaded and unloaded, H_max/M_max ratio of −15° was significantly larger than those of 0° and +15° conditions. The H_max/M_max ratios were 51 %—43%—41% with loaded and 56%—46%—44% with unloaded, for −15° — 0° — +15° slope conditions respectively. Thus, limb loading/unloading had limited impact on the extent of influence that ankle angles exert on the H-reflex excitability. This suggests that task-dependent CNS control of reflex excitability may regulate the influence of sensory input on the spinal reflex during standing.

Introduction

The soleus H-reflex is task and posture dependently modulated (Capaday and Stein, 1986; Llewellyn, Yang and Prochazka, 1990; Hayashi, Tako, Tokuda, 1992; Trimble, Brunt, Jeon, Kim, 2001; Zehr, Collins, Frigon and Hoogenboom, 2003). Factors such as the excitability of the α-motor neuron pool (Angel and Hofmann, 1963; Taborikova and Sax, 1968), joint angle and motion (Capaday and Stein, 1986, 1987; Crenna and Frigo, 1987; Hwang, 2002; Knikou and Rymer, 2002), muscle length (Dyhre-Poulsen, Simonsen and Voigt, 1991; Cheng, Brooke, Misiaszek and Staines, 1995), muscle activation history (Gregory, Wise, Wood, Prochazka and Proske 1998), and activation of other sensory afferents (Zehr, Frigon, Hoogenboom and Collins, 2004; Knikou, 2006) likely contribute to the H-reflex modulation. Known mechanisms of H-reflex modulation includes presynaptic inhibition of Ia afferents (Morin, Pierrot-Deseilligny and Hultborn, 1984; Roby-Brami and Bussel, 1990; Morita, Crone, Christenhuis, Peterson and Nielsen 2001), reciprocal inhibition (Shindo, Harayama, Kondo, Yanagisawa and Tanaka, 1984; Crone, Hultborn, Jesperson and Nielsen, 1987; Nielsen and Kagamihara, 1992; Morita, Shindo Momoi, Yanagawa, Ikeda and Yanagisawa 2006), recurrent inhibition (Hultborn, Lindstorm and Wigstorm 1979; Hultborn and Pierrot-Deseilligny, 1979), and homosynaptic post-activation depression (Crone and Nielsen, 1989; Hultborn, Illert, Nielsen, Paul and Ballengaard 1996), and/or interaction among them (Rudomin, 1999; Enriquez-Denton, Morita, Christensen, Petersen, Sinkjaer and Nielsen, 2002), together with the descending control (Burke, Grades, Mazevet, Meunier and Pierrot-Deseilligny 1992; Iles and Pisini, 1992; Petersen, Christensen and Nielsen, 1998). Presumably, which factors and mechanisms influence the reflex excitability more would depend on the task being performed.

In people lying in prone, H-reflex modulation is related to the muscle spindle afferent discharge that reflect the preceding change in muscle length (Pinniger, Nordlund, Steele and Cresswell, 2001). In people sitting, Frigon et al. (Frigon, Carroll, Jones, Zehr and Collins, 2007) showed that joint angle and background muscle contraction influence the H-reflex/M-wave recruitment curve. In contrast, in people during walking, the robust suppression of the swing-phase H-reflex is present regardless of the extents of knee flexion or hip rotation, or the antagonist tibialis anterior (TA) activation (Yang and Whelan, 1993). The effects of joint angle on H-reflex modulation during a static functional task of standing is currently not understood.

This study aimed to examine the effects of joint angles and limb loading on the H-reflex during standing. We investigated the effects of three different imposed ankle angles on the soleus H-reflex recruitment curve during standing, with the test leg loaded (i.e., weight symmetrically borne with both legs) and unloaded (i.e., full body weight solely borne with the contralateral leg). The unloaded condition was to examine the effects of imposed ankle angles with the test leg at rest (i.e., with the minimum soleus and TA EMG activity), in comparison to the leg in an active task of standing (i.e., loaded). For both the test leg loaded and unloaded conditions, the test leg was in full contact with the sloped (or not sloped) surface, with the knee and hip angles maintained at those of natural standing (i.e., naturally extended). The experimental setup, thus, aimed to minimize the differences in the influence of hip and knee joint afferents between the loaded and unloaded conditions and across the three different slope angles.

Materials and Methods

General Procedure

Fourteen individuals with no known neurological injuries were recruited into the study. For two of them, the maximum H-reflex (H_max) during natural standing was less than 10% of the maximum M-wave (M_max), too small to adequately assess its modulation. Since evaluating the suppression or facilitation of such small H-reflexes would be more error prone (e.g., the H-reflex can never be <0% M_max), we did not include those individuals’ data for further analysis.

The study consisted of two experiments. In experiment 1, all the measurements were made while the participant stood on the sloped (or not sloped) surface with his/her body weight borne with both legs. Thus, the test leg (as well as the contralateral leg) was ‘loaded.” In experiment 2, all the measurements were made while the participant stood with all of his/her body weight borne in the leg that was not tested. This made the test leg ‘unloaded’ while the foot of the test leg was placed on the sloped (or not sloped) surface with the minimum soleus and TA background EMG activity. With no active contraction of the dorsi- or plantarflexor muscles, the test leg’s ankle joint angle was passively altered by the slope of the surface. The test leg’s hip and knee angles remained at or very close to those during natural standing in each participant.

Twelve individuals (7 females and 5males, aged 31.3±6.7 years) were studied in the experiment 1 (i.e., test leg was loaded) and seven individuals (4 females and 3 males, aged 30.1±6.7 years) were studied in the experiment 2 (i.e., test leg was unloaded). The study protocol was approved by the Institutional Review Board of Medical University of South Carolina, Charleston, SC, and all participants gave written consent prior to participating in the study.

EMG signals were recorded from the soleus and TA using self-adhesive surface Ag-AgCl electrodes (2.2 x 3.5 cm; Vermed, Bellows Falls, VT). Soleus electrodes were placed longitudinally just below the gastrocnemius muscles with the centers of electrodes ~3 cm apart. The TA electrodes were placed on the center of muscle belly. EMG signals were amplified, bandpass filtered at 10–1000 Hz, and sampled at 3200 Hz. Ongoing (i.e., background) soleus EMG (BG_EMG) activity was continuously displayed on a computer monitor to help the participant maintain the BG_EMG activity within the specific window (Thompson, Chen and Wolpaw, 2009). When the soleus and TA BG_EMG was kept in the pre-specified range for 2 s, and if 5 s had passed since the last stimulus, a 1-ms square pulse was delivered to the tibial nerve in the popliteal fossa, using a Grass S48 stimulator with a CCU1 constant current unit and a SIU5 stimulus isolation unit (Natus Neurology - Grass, West Warwick, RI), through self-adhesive surface Ag-AgCl electrodes (2.2 × 2.2 cm for the cathode and 2.2 × 3.5 for the anode; Vermed). Soleus BG_EMG was aimed at the natural standing level (typically corresponding to 10–15% of the maximum voluntary contraction level (Thompson et al., 2009)) for the experiment 1, and the resting level (<8μV) for the experiment 2. TA BG_EMG was maintained at resting level (<8μV) for both experiments.

The soleus H-reflex/M-wave recruitment curves were obtained while the participant stood on +15°, −15°, or 0° slope, which aimed to produce the corresponding degrees of ankle joint angles for the standard definition of ankle joint angle, see (Wu, Siegler, Allard, Kirtley,Leardini, Rosenbaum, Whittle and D’Lima, 2002) (see Figure 1A). For each slope condition, the tibial nerve stimulus intensity was increased from below the H-reflex threshold to just above the level that was required to elicit the M_max, in steps of 1.25 mA (Thompson, Chen and Wolpaw, 2013). Four EMG responses were averaged at each stimulus level. The order of three slope conditions was randomized across individual participants. The BG_EMG activity and M-wave size (i.e., value in %M_max) was matched across the three slope conditions. Throughout the experiment, the ankle, knee, and hip joint angles were recorded using electro goniometers (Biometrics Ltd., Newport, UK). Two-dimensional joint angles in the sagittal plane (i.e., dorsiflexion/plantarflexion for the ankle, flexion/extension for the knee and hip) were calculated for the 200 ms pre-stimulus period.

Figure 1.

Experimental slope conditions and typical examples of EMG recording (A): A schematic diagram of a participant standing on three different slope conditions. (B and C): Peristimulus soleus EMG sweeps for the H-reflex/M-wave recruitment curve measurements with the test leg loaded (B) and the test leg unloaded (C) from a representative participant. The H_max and M_max amplitudes differ clearly between −15° (left) and 0° (middle) slope conditions with both the test leg loaded and unloaded.

Data Analysis

The H-reflex and the M-wave amplitudes were measured as peak-to-peak values in time windows determined for each participant. Typically, a time window of 5-22 ms post-stimulus was used for the M-wave and 32-45 ms post-stimulus for the H-reflex. To measure the BG_EMG activity, rectified EMG activity in the 50-ms pre-stimulus period was averaged. To compare across the three slope conditions, Soleus BG_EMG was normalized to the mean rectified amplitude of the M_max calculated in the same time window as for the peak-to-peak value. For TA BG_EMG, absolute values (in μV) were compared across the three slope conditions, since it remained within a resting level of <8μV across the conditions. Then, the stability of BG_EMG activity across the conditions was assessed using random effects analysis of variance (ANOVA). For comparison across the slope conditions, H-reflex and M-wave values are normalized to the M_max measured for each respective slope condition, and a random effects ANOVA was applied. The α level was set at 0.05. Post hoc pairwise comparisons were performed using a Tukey adjustment. Unless otherwise stated, descriptive statistics were reported for the dependent variables (mean ± standard error). All statistical analyses were performed using SPSS® version 24.0.

Results

Ankle, knee, and hip joint angles were successfully imposed across three slope conditions

Effects of different slope conditions (i.e., +15°, −15°, and 0°) were apparent on the ankle joint angles and not on the knee or hip angles. For +15° and −15° slope conditions, the joint angles were calculated as differences from the angles in the 0° neutral slope condition. For the experiment 1 (i.e., the test leg loaded), the ankle angles were −15±1° for the −15° slope condition and were +15±1° for the +15° slope condition, while both the knee and hip angles deviated within ±1° range of the 0° neutral condition. Similarly, for the experiment 2 (i.e., the test leg unloaded), the ankle angles were −15±0° for the −15° slope condition and were +15±0° for the +15° slope condition, while the knee and hip angles deviated within ±2° range of the 0° neutral condition. Thus, these data confirmed that three different slope conditions successfully imposed intended joint angle differences at the ankle.

BG_EMG of soleus and TA across three slope conditions were consistent

For both the experiment 1 and experiment 2, the soleus BG_EMG was maintained across the three slope conditions. For the experiment 1 (i.e., the test leg loaded), the soleus BG_EMG amplitudes were 24±3 μV (1.0±0.3 %M_max), 22±3μV (0.9±0.4 %M_max), and 19±2 μV (0.9±0.3 %M_max) for −15°, 0°, and +15° slope conditions. A random effects ANOVA showed that the soleus BG_EMG (%M_max) amplitude did not differ significantly across all the three slope conditions (p=0.54). Similarly, for the experiment 2 (i.e., the test leg unloaded) the soleus BG_EMG was 6±0 μV (0.4±0.1 %M_max) for all three slope conditions, clearly with no difference among the conditions (p=0.45, a random effects ANOVA).

TA BG_EMG remained at the resting level (i.e., <8μV) across the three slope conditions for both experiments. A random effects ANOVA confirmed that there were no effects of slope condition on the TA BG_EMG amplitudes for both the experiment 1 (p=0.32) and the experiment 2 (p=0.78).

M_max and H_max in the leg loaded (Experiment 1) and unloaded (Experiment 2)

Figure 1B shows typical examples of peri-stimulus EMG sweeps during the H-reflex/M-wave recruitment curve measurements for all 3 slope conditions of the experiment 1. From these superimposed sweeps, the M_max and H_max are easily discemable for each condition. Similar to the previous report by Frigon (Frigon et al., 2007), the shape of M_max appeared to differ across the conditions. The raw values of M_max and H_max, and H_max/M_max ratio for this experiment are summarized in Table 1, together with the random effects ANOVA and post hoc analysis results. For both the M_max and H_max, the amplitudes decreased from the −15° to 0° to +15° slope condition. The H_max/M_max ratio varied significantly within a 12±3(SE) %M_max range across the conditions. H_max/M_max ratios for individual participants across the three slope conditions are shown in Figure 2A.

Table 1.

Summary of the soleus M_max and H_max values

Test Leg Condition	Parameters	ANOVA results	−15° slope (plantarflexion)	0° slope (neutral)	+15° slope (dorsiflexion)
Exp. 1 loaded (n=12)	Mmax (mV)	P<0.0001	10.3 ± 1.3 ^	8.9 ± 1.3 ‡	7.3 ± 1.0 *
	Hmax (mV)	P<0.0001	5.3 ± 0.9 ^	4.0 ± 0.7 ‡	3.0 ± 0.6 *
	Hmax/Mmax	P=0.0002	50.1 ± 4.4 ^	42.6 ± 3.6	40.7 ± 4.5 *
Exp. 2 unloaded (n=7)	Mmax (mV)	P<0.0001	10.3± 1.2 ^	8.7 ± 1.5 ‡	8.3± 1.7 *
	Hmax (mV)	P<0.0001	5.8 ± 0.7 ^	3.9 ± 0.8	3.6 ± 0.9 *
	Hmax/Mmax	P=0.0002	56.4 ± 2.7 ^	46.1 ± 4.6	44.2 ± 5.4 *

Group mean ± standard for H_max (mV), M_max (mV), and H_max/M_max ratio across the three slope conditions. Significant differences (by ANOVA and Tukey test as post hoc) between the conditions are indicated as follows:

^{^}for −15° > 0 °

^‡for 0° > +15 °

^*for −15° > +15 °.

Figure 2.

Normalized H_max amplitudes (expressed as %M_max) in individual participants across the three slope conditions. (A): Values during the test leg loaded. (B): Values during the test leg unloaded.

Typical examples of EMG sweeps for the H-reflex/M-wave recruitment curve measurements for the experiment 2 are shown in Figure 1C. Values of the M_max and H_max, and H_max/M_max ratio are summarized in Table 1, together with the random effects ANOVA and post hoc analysis results. Similar to the trends of the experiment 1, the M_max and H_max amplitudes decreased from the −15° to 0° to +15° slope condition. The H_max/M_max ratio varied significantly within a 13±3(SE) %M_max range across the conditions. Individual H_max/M_max ratios across these slope conditions are shown in Figure 2B.

Discussion

This study found that (1) the M_max and H_max are affected by slope angles, and (2) the extent of H_max/M_max ratio change across slope angles was similar between the test leg unloaded (i.e., not supporting weight) and loaded (i.e., supporting about half of the body weight). Since the soleus and TA BG_EMG and the knee and hip joint angles were maintained across the three slope conditions, observed differences in the M_max and H_max would largely be attributed to different ankle angles imposed by different slope angles.

Effects of ankle angles on the H_max and M_max during standing

In this study, not only the raw amplitudes (i.e., mV) of M_max and H_max but also the H_max/M_max ratio was greater when the ankle was more plantarflexed and the ratio was smaller when the ankle was more dorsiflexed. This finding is in line with an earlier study (Frigon et al., 2007) that investigated the soleus H-reflex/M-wave recruitment curves in seated participants with different ankle joint angles and background EMG levels and found robust influence of those factors on the soleus M_max, H_max, and the H_max/M_max ratio. In the present study, all the measurements were made in stable static, standing postures, as no joint movements occurred in >2s of prestimulus period (see Methods). In such static conditions with the ankle angle being the only joint angle differed among the conditions, it is probable that the plantarflexor muscle length differed among the conditions and affected the measured H-reflex amplitude (Frigon et al., 2007; Dutt-Mazumder, Slobounov, Challis and Newell, 2016). The fact that the M_max and H_max amplitudes changed in the same way (although not to the same extent) further support this possibility. In addition to the proprioceptors in muscles that sense the muscle length, proprioceptors in the ankle joint and cutaneous afferents in foot would also fire differently when the ankle position changes (Cheng et al., 1995; Misiaszek, Cheng and Brooke, 1995; Brooke, Cheng, Collins, McIllroy, Misiaszek and Staines, 1997; Rossi and Mazzocchio, 1988; Burke, Dickson and Skuse, 1991; Iles 1996; Brooke, McIlroy, Staines, Angerilli and Peritore, 1999), and probably contributed to H-reflex amplitude differences between different ankle angles.

A notable finding is that with the test leg loaded or unloaded, the influence of ankle angles was clearly present but appeared to be small; the H_max/M_max ratio varied within 12–13% from the plantarflexed condition to the dorsiflexed condition. Since the H-reflex normalization methods, the angles of proximal joints (i.e., knee and hip), and the examined ranges of ankle angles differ between the studies, the extent of changes cannot be directly compared to Frigon et al. (Frigon et al., 2007). However, since the soleus H-reflex is known to be modulated posture and/or position dependently from lying supine or prone to semi-reclined (seated) to standing (Hayashi et al., 1992; Koceja, Markus and Trimble, 1995; Angulo-Kinzler, Mynark and Koceja, 1998; Goulart, Valls-Sole and Alvarez, 2000; Chalmers and Knutzen, 2002), it is quite possible that the extent to which the joint angle influences the H-reflex excitability could differ between postures (Brooke et al., 1997; Zehr and Stein, 1999). It should also be noted that electrode placements, particularly for the soleus, may differ between the present study and Frigon et al., (Frigon et al., 2007) and Frigon et al. data show that placement affects magnitude of EMG activity.

Limb loading effects on the H-reflex excitability during standing

The present study asked whether the limb loading or unloading can affect the extent to which the joint angle influences the H-reflex excitability during standing. On the 0° slope (i.e., neutral) surface, the average H_max/M_max ratio was slightly higher (by ≈3% M_max) with the leg unloaded than loaded. This small, statistically non-significant difference only moderately agrees with a previous study by Nakazawa et al. (Nakazawa, Miyoshi, Sekiguchi, Nozaki, Akai, and Yano, 2004), who showed that ≈20N of ankle joint unloading increases the H_max/M_max ratio by ≈10%M_max during standing. With a set of unique water-tank experiments, Nakazawa et al. manipulated the amount of loading at the ankle joint while maintaining the same small amount of body weight borne (20% of body weight) and the same small amount of soleus EMG activity (21-25% of standing level on the ground), and delineated ankle joint loading and unloading effects on the soleus H-reflex. The difference between the present results and the results by Nakazawa et al. suggests that limb loading and isolated joint loading affect the excitability of spinal reflex pathways differently or to different extents, although many of the same sensory afferents could be involved in both loading conditions.

In the present sets of experiments, the change of H_max/M_max ratio from the plantarflexed condition to the dorsiflexed condition was similar between the test leg unloaded (13%) and loaded (12%). This was unexpected, since the experimental procedures ensured that the amount of soleus BG_EMG would differ between the conditions; indeed, the unloaded condition resulted in resting level (i.e., <8 μV), whereas the loaded condition resulted in the active soleus contraction (i.e., ≈20 μV). It has been suggested that inhibitory mechanisms acting on spinal reflexes are reduced in actively contracting muscles, and thus, may reduce the differences in the measured reflex amplitude between experimental conditions or between participant populations (Nielsen, Petersen, Crone and Sinkjaer, 2005; Stein and Thompson, 2006; Dietz and Sinkjaer, 2012). Frigon et al., (Frigon et al., 2007) also found similar effects of BG_EMG on how much the ankle angle influences the H-reflex excitability; in their seated participants, the difference in the soleus H-reflex between ankle angles progressively diminished as the background EMG increased. Thus, the active muscle contraction would have caused at least some difference in spinal inhibition. Further, limb loading is known to produce critical signals for regulating locomotion (Duysens and Pearson, 1980; Duysens, Clarac and Cruse, 2000; Pearson and Collins, 1993; Hiebert, Whelan, Prochaska and Pearson, 1996; Whelan, 1996; Stephens and Yang, 1999; Donelan and Pearson, 2004) and is essential for improving locomotion after spinal cord injury and other CNS disorders (Dietz, 2002; Dietz, Muller and Colombo, 2002). Loading modulates important spinal reflexes, such as Ib inhibition (Faist, Hoefer, Hodapp, Dietz, Berger and Duysens, 2006), stretch reflexes (Sinkjaer, Andersen, Ladouceur, Christensen and Nielsen, 2000; Grey, van Doornik and Sinkjaer, 2002), and cutaneous reflexes (Bastiaanse, Duysens and Dietz, 2000; Duysens, Tripple, Hortstmann and Dietz, 1990), and interneurons that receive inputs from load receptors interact with sensory afferents arising from joint receptors and other proprioceptors (including muscle spindle afferents) (Brooke et al., 1997; Staines, Brooke, Angerilli and McIllroy, 1998; Duysens, Bastiaanse and Smits-Engelsman, 2004). Then, why were there little loading effects in the present study?

A plausible explanation is that the excitability of such neuronal connections can also be readily modified task-dependently through spinal and supraspinal mechanisms (Duysens et al., 1990; Yang and Stein, 1990; Yang and Whelan, 1993; Pearson, 1995; Stein, 1995; Stephens and Yang, 1999; Zehr and Stein, 1999; Duysens et al., 2004). Thus, it is possible that differences in the firing of different sensory afferents (e.g., load receptors, joint receptors, spindle afferents) between the loaded and unloaded conditions were minimized by the task-dependent control of spinal reflexes (Yang and Whelan, 1993). In fact, during an important task of walking, unloading and reduced EMG activity have little effects on phase-dependent modulation of the soleus H-reflex excitability (Knikou, Hajela, Mummidisetty, Xiao and Smith, 2011). Together with a small range of H_max/M_max ratios found across different slope (i.e., ankle) angles (see Figure 2), a lack of limb loading effects on the extent of H_max/M_max ratio modulation may illuminate the fact that standing is a significant motor task of its own. The present results support the view that the sensory afferents that sense the ankle angle and affect the H-reflex excitability are susceptible to the spinal and supraspinal controls that produce task-dependent modulation of spinal reflexes (Stein and Capaday, 1988; Stein, 1995; Stein and Thompson, 2006).

Conclusion

In the present study, we found that the soleus H_max/M_max ratio changes across different imposed ankle angles during standing, and the extent of H_max/M_max ratio change was similarly small (12-13% M_max) with the test leg loaded and unloaded. The present study shares both differences and similarities with the previous studies; by removing certain factors from a functional task (e.g., changing the joint angle or muscle activity in sitting, and loading a single joint without loading the whole limb), the previous studies revealed specific effects of background muscle activity and joint angles (Frigon et al., 2007) or joint loading (Nakazawa et al., 2004) on reflex excitability. Those studies provide fundamental information on how different sensory afferents may behave, interact with each other, and influence the excitability of a reflex pathway, in experimentally simplified settings. Only with such knowledge, we can begin to understand the mechanisms of coordinated excitation and inhibition of multiple muscles that serve a motor function. By examining the effects of ankle angles with or without limb loading in standing, this study, in combination with those previous studies, may provide insight into complex neurophysiology of a simple yet significant motor task of standing.